دانلود مقاله ISI انگلیسی شماره 39136
عنوان فارسی مقاله

اولین گام در استفاده از یادگیری ماشین در داده های fMRI برای پیش بینی خاطرات مزاحم از بخش های ضبط شده فیلم پس از سانحه

کد مقاله سال انتشار مقاله انگلیسی ترجمه فارسی تعداد کلمات
39136 2014 10 صفحه PDF سفارش دهید محاسبه نشده
خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.
عنوان انگلیسی
First steps in using machine learning on fMRI data to predict intrusive memories of traumatic film footage
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Behaviour Research and Therapy, Volume 62, November 2014, Pages 37–46

کلمات کلیدی
خاطرات نفوذی - تروما - یادگیری ماشین - فلاش بک - MVPA تصویربرداری رزونانس مغناطیسی عملکردی - تصویرسازی ذهنی
پیش نمایش مقاله
پیش نمایش مقاله اولین گام در استفاده از یادگیری ماشین در داده های fMRI برای پیش بینی خاطرات مزاحم از بخش های ضبط شده فیلم پس از سانحه

چکیده انگلیسی

Abstract After psychological trauma, why do some only some parts of the traumatic event return as intrusive memories while others do not? Intrusive memories are key to cognitive behavioural treatment for post-traumatic stress disorder, and an aetiological understanding is warranted. We present here analyses using multivariate pattern analysis (MVPA) and a machine learning classifier to investigate whether peri-traumatic brain activation was able to predict later intrusive memories (i.e. before they had happened). To provide a methodological basis for understanding the context of the current results, we first show how functional magnetic resonance imaging (fMRI) during an experimental analogue of trauma (a trauma film) via a prospective event-related design was able to capture an individual's later intrusive memories. Results showed widespread increases in brain activation at encoding when viewing a scene in the scanner that would later return as an intrusive memory in the real world. These fMRI results were replicated in a second study. While traditional mass univariate regression analysis highlighted an association between brain processing and symptomatology, this is not the same as prediction. Using MVPA and a machine learning classifier, it was possible to predict later intrusive memories across participants with 68% accuracy, and within a participant with 97% accuracy; i.e. the classifier could identify out of multiple scenes those that would later return as an intrusive memory. We also report here brain networks key in intrusive memory prediction. MVPA opens the possibility of decoding brain activity to reconstruct idiosyncratic cognitive events with relevance to understanding and predicting mental health symptoms.

مقدمه انگلیسی

Introduction The focus of the current paper is on using neuroimaging to understand the development of intrusive memories of trauma, that is “recurrent, involuntary and intrusive distressing memories of the traumatic event” (The Diagnostic and Statistical Manual of Mental Disorders, 5th ed.; DSM-5; American Psychiatric Association, 2013). Intrusive memories are a hallmark symptom from the re-experiencing cluster of Post-Traumatic Stress Disorder (PTSD). They have previously been defined as involuntary mental images that occur in a waking state (Frankel, 1994 and Jones et al., 2003). Thus, key features of intrusive memories are that they are involuntary rather than deliberately retrieved, i.e. apparently spontaneous (Kvavilashvili, 2014); include perceptual aspects of the traumatic event, i.e. involve mental imagery rather than only verbal thought (Holmes, Grey, & Young, 2005); are in line with episodic and memory recall more broadly (Conway, 2001), and have distressing, i.e. emotional content (Hackmann, Ehlers, Speckens, & Clark, 2004). For example, after a motor vehicle accident, seeing scaffolding smashing through the car windscreen (see Grey and Holmes, 2008 and Holmes et al., 2005 for further examples). In their most extreme form, re-experiencing symptoms can present as so-called dissociative ‘flashbacks’ where patients relive past events as if they are happening in the present (American Psychiatric Association, 2013). In contrast, during the experience of an intrusive memory the past events are spontaneously remembered while awareness of the present is maintained. Due to the nature of this special issue, “How neuroscience informs behavioural treatment” within Behaviour Research and Therapy, we appreciate that many readers may not have a detailed understanding of neuroimaging terms and techniques. We therefore present a slightly longer than normal introduction to guide the reader through the steps taken before performing the main predictive analysis presented here. We first describe our initial study using traditional neuroimaging analysis techniques ( Bourne, Mackay, & Holmes, 2013) and its subsequent replication ( Clark, Holmes, Woolrich, & Mackay, submitted for publication). We then introduce the ideas of multivariate pattern analysis (MVPA) and machine learning, before next describing how we utilised these techniques in the current experiment. The aim of this is to provide a methodological basis for understanding the context of the current results and show that these findings are both replicable and reliable. We believe that by using neuroimaging techniques in addition to behavioural, cognitive and psychophysiological experiments we may be able to identify those neural and cognitive functions that are critical for intrusive memory formation. Understanding how intrusive memories are formed from multiple perspectives may allow future work to improve the ability to refine treatments which target the underlying mechanisms of intrusive memory (i.e. symptom) development. Indeed, by gaining the most comprehensive understanding of differences at an individual level, we may be able to open future possibilities of early screening for risk of PTSD, as well as the development of preventative approaches in the immediate aftermath of trauma and for targeted early interventions. We also note that many different approaches to machine learning and MVPA are evolving, including (but not limited to) Random Forest Theory (Breiman, 2001), Graph theory (Power et al., 2011 and Sporns, 2014) and Representational Similarity Analysis (Kriegeskorte, Mur, & Bandettini, 2008), in addition to that used here, a Support Vector Machine classifier (Pereira, Mitchell, & Botvinick, 2009). The current work represents only first steps in applying neuroimaging techniques to understand the neural impact of witnessing trauma and to inform behavioural treatment. We finish by exploring how such techniques might have implications for future cognitive behavioural therapy. Intrusive memories and PTSD Most people will experience a traumatic event during the course of their lifetime and a significant minority will go on to develop PTSD (Breslau et al., 1998 and Kessler et al., 1995). We have successful treatments for the full blown disorder, those recommended by clinical guidelines (e.g. National Institute for Health and Clinical Excellence, 2005) are Cognitive Behavioural Therapy (CBT; e.g. Ehlers and Clark, 2000 and Foa and Rothbaum, 1998) and Eye Movement Desensitisation and Reprocessing (EMDR; Shapiro, 1995). However, satisfactory preventative treatments against PTSD development are lacking (Roberts, Kitchiner, Kenardy, & Bisson, 2009). A greater understanding of the brain mechanisms that lead to the development of intrusive memories may help guide effective preventative interventions for the early aftermath of trauma. We know little, in particular in terms of neuroscience, about why only certain events within a trauma return as intrusive memories when others do not. Processing at the time of trauma (peri-traumatic) is implicated in PTSD development (e.g. Brewin, 2014, Ehlers and Clark, 2000 and Ozer et al., 2003). Additionally, experimental findings implicate heightened emotional processing at the time of the event in intrusive memory development (Clark et al., 2013 and Clark et al., 2014). Interestingly, dissociation, defined within the DSM 5 as “a disruption of and/or discontinuity in the normal integration of consciousness, memory, identity, emotion …” (American Psychiatric Association, 2013, p. 291), can be a reaction to extreme emotion, and peri-traumatic dissociation has also been associated with intrusive memory formation (e.g. Daniels et al., 2012 and Holmes et al., 2004). Seminal work on ‘flashbulb’ memories, defined as ‘memories for the circumstances in which one first learned of a very surprising and consequential (or emotionally arousing) event’ (Brown & Kulik, 1977) may also illuminate some of the mechanisms involved in intrusive memory formation. While flashbulb memories are a distinct phenomenon (and not exclusive to trauma, but part of autobiographical memory more generally), they may lie on a continuum with intrusive memories. Research suggests that memories that end up as flashbulb memories are psychophysiologically arousing, personally salient and unexpected and sudden (Brown & Kulik, 1977). Indeed, psychophysiology has been associated with intrusive memory development; at the time of viewing a specific film scene that is later recalled as an intrusive memory, heart rate has been shown to drop in comparison to the rest of film viewing (Chou et al., 2014 and Holmes et al., 2004). Understanding the neural processes involved in intrusive memory formation adds another level of comprehension of this complex phenomenon. Neuroimaging and established PTSD The majority of studies using neuroimaging to investigate PTSD have done so once symptoms are already established in patients (Francati et al., 2007, Hughes and Shin, 2011 and Pitman et al., 2012). Neurocircuitry models suggest that PTSD is characterised by reduced activity in the ventromedial prefrontal cortex, which is associated with decision making and emotional response inhibition, and increased activation in the amygdala and other limbic areas, which are associated with emotional processing (e.g. Rauch et al., 2006 and Rauch et al., 1998). A further recent model suggests that abnormalities in the amygdala and dorsal anterior cingulate cortex are pre-disposing, while abnormal interactions between the hippocampus and ventromedial prefrontal cortex arise after developing PTSD (Admon, Milad, & Hendler, 2013). While informative for understanding PTSD as a whole, these studies cannot tell us specifically about intrusive memories, that is, those events we need to target within a CBT treatment (e.g. Ehlers and Clark, 2000 and Foa et al., 2007). Further, studying symptoms once they are already established tells us little about the neural processes involved in intrusive memory formation (aetiology). The trauma film paradigm: an experimental psychopathology approach Electronic media offers a way to use neuroimaging to investigate the brain responses to experimental analogue trauma exposure and intrusive memory formation. Recent work has examined the effects of electronic media, for example television news film footage, on the development of PTSD symptoms. Individuals exposed for prolonged hours to media footage of terrorist attacks have been shown to present higher scores on stress and trauma related symptom scales both a month after the attack (Holman, Garfin, & Silver, 2014) and 2–3 years after the attack (Silver et al., 2013). Additionally, the DSM 5 (American Psychiatric Association, 2013) now includes exposure to trauma through electronic media in the definition of a traumatic event, with the caveat that the exposure is work related. Together, this suggests that traumatic events transmitted through electronic media footage have the potential to induce PTSD-like symptomatology. The trauma film paradigm is widely used as an experimental analogue of psychological trauma (see Holmes and Bourne, 2008 and Lazarus, 1964) and involves healthy participants viewing traumatic footage in line with DSM 5 criteria for a traumatic event (e.g. real life footage depicting actual or threatened death and serious injury; American Psychiatric Association, 2013). The paradigm has been most commonly used in behavioural experiments. Examples include the investigation of cognitive tasks to reduce intrusive memory frequency (e.g. Tetris; Holmes, James, Coode-Bate, & Deeprose, 2009) or vulnerability factors for intrusive memory development (Laposa and Alden, 2008 and Wessel et al., 2008). Recently, we conducted the first study, to our knowledge, to combine the trauma film paradigm with functional Magnetic Resonance Imaging (fMRI) (Bourne et al., 2013; n = 22). This provided a prospective measure of the brain activation at the moment of viewing a film scene that would later return as an intrusive memory during the following week. We then replicated this experiment, finding a near identical pattern of results ( Clark et al., submitted for publication; n = 35). The importance of such replication studies has been particularly noted recently within the field of fMRI (e.g. Carp, 2013 and Fletcher and Grafton, 2013). In these studies, unlike most fMRI designs, we could not specify our neuroimaging ‘events’ of interest in advance (i.e. the specific time within stimuli presentation when brain activation is selected to be compared to the rest of stimuli presentation). This is due to intrusive memories being highly idiosyncratic; thus we did not know which scenes in the film would return involuntarily for each individual (just as after a real trauma we do not know which moments will be the hotspots and intrude). The film was created to include 20 scenes that had previously been found to induce intrusive memories. Participants recorded their intrusive memories (defined as mental images of the film content that involuntarily come to mind) for one week in daily life using a pen-and-paper diary. From written descriptions in the intrusive memory diary, intrusions were matched to specific scenes within the film (e.g. the car rolling over the hedge hitting the boy playing football in his garden). Film scenes were then classified on an individual participant basis as either ‘Flashback scenes’ – emotional scenes that returned as an intrusive memory for that individual, or ‘Potential scenes’ – emotional scenes that did not return as an intrusive memory for that individual, but did in other participants (see Fig. 1). On average, 3 of the possible 20 scenes became intrusive memories for each participant; a similar frequency to the number of different events experienced as intrusions after real life trauma (Grey and Holmes, 2008 and Holmes et al., 2005). Procedure diagram. Participants viewed traumatic footage while undergoing fMRI. ... Fig. 1. Procedure diagram. Participants viewed traumatic footage while undergoing fMRI. Specific scenes in the film were determined to be ‘Possible’ scenes (scenes that had previously caused intrusive memories in other studies). As intrusive memories are idiosyncratic, Possible scenes became either ‘Flashback’ scenes or ‘Potential’ scenes for each individual. Scene type was determined for each participant retrospectively from the 1 week intrusive memory diaries. Figure options Using a standard statistical mass univariate regression analysis approach (i.e. the analysis currently most used for fMRI data) we found that Flashback scenes, in comparison to Potential scenes, were characterised by widespread increases in brain activity including the anterior cingulate cortex, thalamus, putamen, insula, amygdala, ventral occipital cortex, left inferior frontal gyrus and bilateral middle temporal gyrus. In brief, brain regions that have previously been associated with emotional processing, visual/mental imagery and memory (see Bourne et al., 2013 for discussion). These results provided, to our knowledge, the first evidence of a ‘neural signature’ at the time of intrusive memory formation. Predicting from fMRI; multivariate pattern analysis (MVPA) and machine learning However, traditional univariate fMRI analysis only highlights an association of peri-traumatic brain responses with later intrusive memories across a group of individuals (see for details Jezzard et al., 2001 and Smith et al., 2004). Additionally, traditional fMRI analysis relies on the self-report diary to identify the scene type. It would be useful to know the extent to which brain responses during exposure to analogue trauma can actually predict a specific moment of the traumatic footage that would later become an intrusive memory, for example, to inform preventative interventions against intrusive memory formation. Machine learning and multivariate pattern analysis (MVPA) are neuroimaging analysis techniques that can be used to measure prediction accuracy. MVPA makes use of multivariate, spatially extensive patterns of activation across the brain. The patterns of activation across these larger regions can be “learned” through approaches from the field of machine learning. Supervised machine learning techniques optimise input “features” to best separate or describe the two labelled classes of data (i.e. Flashback scene or Potential scene). These “features” are simply summary measures of some aspects of the data. It is through these optimisation steps that machine learning approaches “learn” the patterns that best describe each class of data. Once the patterns have been identified, they can be used to predict the behaviour of new, previously unseen participants. Such approaches can provide greater discriminative ability than spatially localised mass-univariate regression analyses (see for further details, Haxby, 2012, Haynes and Rees, 2006, McIntosh and Mišić, 2013, Mur et al., 2009 and Norman et al., 2006). Machine learning can then be used to learn these patterns of activity to accurately predict the occurrence of a new, unseen example of the same event (Lemm et al., 2011 and Pereira et al., 2009). To highlight just a few examples of MVPA techniques applied to fMRI, neural patterns identified by MVPA while participants were exposed to a shock during the presentation of picture stimuli have predicted the later behavioural expression of fear memory (pupil dilation response) between 2 and 6 weeks after encoding (Visser, Scholte, Beemsterboer, & Kindt, 2013). Additionally, MVPA techniques have identified patterns of activation at encoding that can predict later deliberate memory recall (see Rissman & Wagner, 2012). We hypothesised that machine learning may be able to predict an intrusive memory from just the peri-traumatic brain activation. We aimed first, to investigate whether specific scenes in the film could be identified as later becoming intrusive memories solely from brain activation at the time of viewing traumatic footage by applying machine learning with MVPA. Second, we explore which brain networks are key in MVPA-based prediction of intrusive memory formation, and when the activation of these brain networks in relation to the timing of the intrusive memory scene is important.

نتیجه گیری انگلیسی

Conclusions Using machine learning and MVPA on fMRI data of trauma film encoding, we have demonstrated that peri-traumatic brain activation is able to predict moments that would later return as an intrusive memory with 68% accuracy across participants and within a given participant with 97% accuracy. Here, we make an attempt to import ideas from basic neuroscience to contribute to an area of mental health – intrusive trauma memories. We suggest certain advance neuroimaging techniques may even be developed for use in studying relatively infrequently occurring and idiosyncratic events in mental health symptomatology (such as intrusive memories) and be used to predict individual's future symptom response.

خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.